TOEFL Speaking in 2026: At a Glance
ETS redesigned the TOEFL iBT in January 2026, and the Speaking section changed more than any other part of the test. The four old tasks, with their reading passages, lectures and preparation timers, are gone. In their place are two task types that take about eight minutes in total: Listen and Repeat and Take an Interview.
The section is not face to face. There is no examiner: you listen to the audio, speak into a microphone, and your responses are scored by AI. This guide explains what each task asks you to do, how the section is scored and how to practise, whether you are preparing in India or anywhere else. The format facts come from ETS's own description of the test (TOEFL iBT test content, ets.org).
If you are a parent or a younger sibling starting earlier — students 11+ preparing for English exams on PenLeap often build the structured-answer habits that TOEFL Speaking later rewards, especially for the opinion questions in Take an Interview.
- Items: 11 in total: 7 Listen and Repeat + 4 Take an Interview
- Time: about 8 minutes
- Format: computer-delivered; you speak into a microphone; no live examiner
- Scoring: AI-scored, up to 5 points per item
- Score scale: a Speaking score on the new 1–6 band scale (aligned with CEFR A1 to C2), plus a comparable 0–120 overall score during a two-year transition
- Structure: linear, so the section does not adapt to your answers
- Audio accents: North American, British, New Zealand and Australian
Each task opens with a short scenario introduction, spoken and shown on screen, sometimes with an image, a map or a form. Then the items follow, and your responses are short and timed.
How the Whole TOEFL iBT Changed in 2026
Speaking is only one part of the update. The full test is now shorter, at about an hour and a half of test time (ETS gives 1 hour 23 minutes to 1 hour 29 minutes, and advises allowing about two hours in total), and two of its four sections adapt to your performance.
The TOEFL iBT Sections Since January 2026
| Section | Items | Task types | Time |
|---|---|---|---|
| Reading | 50 (some unscored) | Complete the Words; Read in Daily Life; Read an Academic Passage | about 30 min |
| Listening | 47 | Listen and Choose a Response; Listen to a Conversation; Listen to an Announcement; Listen to an Academic Talk | about 29 min |
| Writing | 12 | Build a Sentence (10); Write an Email (1); Write for an Academic Discussion (1) | about 23 min |
| Speaking | 11 | Listen and Repeat (7); Take an Interview (4) | about 8 min |
- Reading and Listening are adaptive. Each uses a two-stage design: you first take a routing module, and your performance there decides whether your second module is a lower or an upper one. Writing and Speaking are linear.
- Writing changed. The integrated writing task has been removed. Write for an Academic Discussion stays, and Write an Email and Build a Sentence are new. The email and academic discussion responses are scored by AI.
- Scores changed. You get four section scores and an overall score on a 1–6 scale. The overall score is the average of the four sections, rounded to the nearest half band.
How TOEFL Speaking Is Scored Now
Every Speaking response is scored by AI, with up to 5 points per item. Your item scores feed a Speaking section score on the new 1–6 band scale, reported next to your Reading, Listening and Writing scores. The overall score is the average of the four section scores, rounded to the nearest half band, and the bands are aligned with the CEFR levels A1 to C2.
For a two-year transition after January 2026, you also receive a comparable overall score on the 0–120 scale, so you can match it against requirements still written on the old scale. There is no longer a Speaking score out of 30.
What Each Speaking Task Measures
| Task | Items | What ETS says it measures | What that means for your practice |
|---|---|---|---|
| Listen and Repeat | 7 | Processing spoken English and producing accurate, clearly intelligible speech | Catch every word, keep the small words, and say the sentence clearly with natural stress |
| Take an Interview | 4 | Communicating clearly, at a natural pace, with appropriate vocabulary and grammar | Answer the question actually asked, develop it with reasons and examples, and keep a steady pace |
Listen and Repeat (7 Items)
In Listen and Repeat, you listen to short sentences and repeat them exactly as you hear them. The sentences get longer and more complex as the task goes on, so the first few feel easy and the last ones test how much you can hold in memory.
ETS says the task measures how well you process spoken English and produce accurate, clearly intelligible speech. In practice that means two things: you must catch every word, and you must say it back so that a listener, here the scoring engine, can follow it easily.
How to Prepare for Listen and Repeat
1. Listen in chunks, not word by word
Nobody can hold a 20-word sentence as 20 separate words. Fluent listeners hear it as three or four thought groups (chunks), each built around one key word. Train yourself to hear the chunks:
Chunking example
"The writing centre / on the second floor / will be closed / until next Monday."
Four chunks are easy to remember; fourteen loose words are not. When you repeat, pause very briefly at the same chunk boundaries the speaker used.
2. Copy the stress and the melody
English stresses content words (nouns, main verbs, adjectives, adverbs) and squeezes function words. In "I'm GOing to the LIbrary to reTURN some BOOKS", the capitalised syllables are long and clear, while "to the" and "some" are quick and light (/tə ðə/, /səm/). Many Indian speakers give every syllable equal weight, which makes long sentences sound flat and harder to follow. Copy the speaker's rhythm: strong syllables strong, weak syllables weak.
Copy the intonation too: a falling tone at the end of a statement, a rising tone on a yes/no question, and a slight rise on each item in a list until the last one. Our guides to word stress rules and English intonation patterns cover the details.
3. Shadow every day
Shadowing means playing a short recording and speaking along with it, a fraction of a second behind, copying the words, rhythm and tone. It is the closest everyday drill to Listen and Repeat. A simple 15-minute routine:
- Pause and repeat (5 minutes): play one sentence, pause, repeat it exactly, then play it again to check every word.
- Shadow (5 minutes): play the same clip and speak along with it without pausing.
- Stretch (5 minutes): move to longer sentences as the short ones become easy, the same way the test does.
Use clips from campus life and academic talks, such as university podcasts, orientation videos and lecture introductions, and mix accents from North America, the U.K., Australia and New Zealand, since the test audio uses all four.
4. Start promptly, then speak clearly
Begin speaking soon after the sentence ends, before it fades from memory, but do not rush. A clear, evenly paced repeat is easier to understand than a fast one. If you lose part of a long sentence, keep going smoothly with what you caught rather than stopping dead.
Practice drill: 7 sentences, short to long
These practice sentences are written by TalkDrill in a campus setting; they are not taken from the test. Say each one aloud with the chunk breaks (/) and stressed syllables (CAPITALS) shown, then cover it and say it from memory, checking every word.
1. "The BOOKshop OPens at NINE."
2. "Please reTURN your KEYS / to the FRONT DESK."
3. "The biOLogy LAB / has MOVED / to the SECond FLOOR."
4. "If you MISS the SHUTtle, / the NEXT one LEAVES / in TWENty MINutes."
5. "STUdents who WANT a LOCKer / should SIGN UP / at the FRONT DESK / by FRIday."
6. "The LIbrary will stay OPen / until MIDnight / during eXAM week, / but the canTEEN / will CLOSE at TEN."
7. "Before you subMIT your ESsay online, / make SURE / your NAME and STUdent NUMber / are on the FIRST PAGE."
To practise with audio instead of text, TalkDrill's free Echo Chamber game plays a model sentence, you say it back, and it scores how closely and how quickly you matched it.
Take an Interview (4 Questions)
Take an Interview is a simulated interview related to academic or campus situations. You answer four questions about your experiences and opinions. The questions progress from brief factual answers to questions that ask for an opinion, an explanation, a prediction or a narration (a short account of something that happened). ETS says the task measures your ability to communicate clearly, at a natural pace, with appropriate vocabulary and grammar.
The interviewer is part of the test, not a live person, so you cannot ask for a question to be repeated or explained. Listen to the whole question, notice what type it is, and answer that question, not the one you prepared for.
- Early questions: brief factual answers about you, your studies or your routine
- Later questions: opinion, explanation, prediction or narration, which need a developed answer with reasons and examples
Answer Frameworks for Each Question Type
A framework is a shape for your answer, not a memorised script. Learn these five shapes so you can start speaking straight away, then fill them with real details from your own life.
1. Brief factual question: Answer + one detail
Shape: direct answer, then one supporting detail (where, when, how often or why).
Example: "What are you studying?" → "I'm in the second year of a computer science degree. I chose it because I enjoy building small apps in my free time."
Tip: one or two full sentences. Not one word, and not a speech.
2. Experience or narration question: Scene → Problem → Action → Result
Shape: when and where, then the situation or problem, then what you did, then the result or what you learned.
Language: past tenses and sequence words: "Last semester…", "At first…", "Then…", "In the end…".
3. Opinion question: Point → Reason → Example → Point
Shape: a clear position, the main reason, a specific example, then a one-line return to your position. If there is time, acknowledge the other side briefly: "Some students might prefer…, but…".
4. Explanation question: Main idea → How or why → Example → Result
Shape: state the main cause or the first step, explain it with "because" or "so", give an example, then say what it leads to.
5. Prediction question: Prediction → Reason from today → Effect → Hedge
Shape: what you think will happen, why (based on what you see now), what it will change, and a sensible hedge such as "probably", "I expect" or "it might".
Language: "will", "is likely to", "might", "in the next few years".
Sample Interview Questions and Answers
The questions and answers below are written by TalkDrill to show the frameworks in action. They are not official ETS items, and we do not attach scores to them, because the test's AI scoring cannot be reproduced outside the test.
Scenario: A researcher is interviewing students about campus life
Q1 (brief factual): "How do you usually get to your classes?"
"I usually walk. My hostel is about ten minutes from the main building, and walking gives me time to wake up before my first lecture."
Q2 (experience and narration): "Tell me about a time a group project did not go as planned."
"Last semester, four of us had to present a survey on student transport. Two days before the deadline, we realised that half our responses came from the same hostel, so the results weren't fair. I suggested we split up and each collect twenty more responses from other parts of the campus. It meant a very long day, but our final data was much more balanced, and the professor mentioned our sampling in her feedback. I learned to check the data early, not at the end."
Q3 (opinion): "Should universities record every lecture and put it online?"
"I think they should, mainly because recordings help students review difficult material. For example, in my statistics course I often understood a topic only on the second listen, when I could pause and take proper notes. Some people worry that students would stop attending, and that's a fair point, but in my experience most students still come to class and use the recordings for revision. So overall, I'd support recording lectures."
Q4 (prediction): "How do you think students will study ten years from now?"
"I expect students will rely much more on AI tools to explain topics and test them, because many students already use them for revision today. That will probably make it easier to study at your own pace. But I think classrooms will still matter, because discussing ideas with other people is hard to replace. So my guess is a mix: more personalised study at home, and more discussion in class."
One more: an explanation answer
Q: "Why do you think many students find their first year at university difficult?"
"I think the main reason is that students suddenly have to manage their own time. At school, teachers check your homework every day, but at university nobody reminds you, so it's easy to fall behind. For example, in my first semester I left most of my reading until the week before exams, and I really struggled. Once I started planning my week every Sunday, things got much easier. So the difficulty is less about the subjects and more about learning to organise yourself."
More practice questions by type
Brief factual: "Where do you usually study?" · "What time does your first class start?" · "Which courses are you taking this term?"
Experience and narration: "Describe a time you asked a teacher for help." · "Tell me about your first week at college." · "Talk about a club or event you took part in."
Opinion: "Should first-year students live on campus?" · "Is it better to take notes by hand or on a laptop?" · "Should attendance count toward grades?"
Explanation: "How do you prepare for a big exam?" · "Explain how you choose which courses to take." · "Why do some students stop going to lectures?"
Prediction: "How will online courses change universities?" · "What will students need most from a library in the future?" · "How do you think your field will change in the next ten years?"
How to Extend an Answer Naturally
A common problem in interview-style speaking is the answer that stops after one sentence. The fix is not to talk faster or add filler, but to add one more layer of meaning. Six reliable layers:
- A reason: "…because…"
- A specific example: "For example, last month…"
- A contrast: "…but…" or "Although…"
- A result: "…so…" or "That's why…"
- A comparison: "At school it was…, but at college…"
- A feeling or a lesson: "What I learned was…"
Q: "Do you prefer studying in the library or in your room?"
Too short: "The library."
Extended: "I prefer the library, because there are fewer distractions. In my room I keep checking my phone, but in the library everyone around me is working, so I focus better. The only time I study in my room is late at night, after the library closes."
Common Mistakes and How to Fix Them
These problems are common when learners practise for the new Speaking section, and several are especially common among speakers of Indian English:
Mistake 1: Preparing with Old-Format Material
Problem: Many books, apps and videos still teach the four-task test, with preparation timers and integrated reading-and-lecture summaries.
Fix: If a resource mentions Task 1 to Task 4, a reading passage before you speak, or a Speaking score out of 30, it describes the old test. Use it only for general speaking practice, and build your routine around Listen and Repeat and Take an Interview.
Mistake 2: Paraphrasing in Listen and Repeat
Problem: Saying the "meaning" of the sentence in your own words, or correcting what sounds like an error.
Fix: The task asks for the sentence exactly as you heard it. Practise covering a sentence and repeating it word for word, then check it.
Mistake 3: Dropping Small Words and Endings
Problem: Leaving out "the" and "a", plural -s and past-tense -ed endings. It is common in Indian English and among many other learners, and in Listen and Repeat every missing piece makes your version different from the original.
Fix: Train the endings on purpose. TalkDrill's Read Aloud passages on past-tense -ed endings and -s and plural endings score every word you read and highlight the ones you missed.
Mistake 4: Wrong Stress on Common Academic Words
Problem: Mispronouncing common academic words like "phenomenon" (stress on NOM), "analysis" (stress on NAL), "determine" (stress on TER) and "specific" (stress on CIF).
Fix: Build a list of 50 common academic words and practise their stress patterns. TalkDrill's word-stress Read Aloud passage scores every word you read and highlights the ones you stressed wrongly.
Mistake 5: Speaking Too Fast
Problem: Rushing, blending words together and dropping syllables to sound fluent. ETS lists a natural pace as part of what Take an Interview measures.
Fix: Record yourself and listen back. Practise at a calm, conversational speed and pronounce every word clearly, especially word endings (-ed, -s, -tion).
Mistake 6: One-Line Answers or Memorised Speeches
Problem: Answering the later interview questions in a single sentence, or reciting a prepared speech that does not quite answer the question.
Fix: Identify the question type, use its framework, and add one layer (a reason, an example or a result) to every answer after the first.
Mistake 7: Overloading with Filler Words
Problem: Excessive "basically", "actually", "you know" and "like" use up time and sound unfocused.
Fix: Replace fillers with brief, purposeful pauses. A one-second pause is much better than saying "basically" three times.
4-Week Preparation Plan for the 2026 Speaking Section
Week 1: Learn the Format and Build the Habits (Days 1–7)
- Read ETS's description of the Speaking section and look for official practice material on ets.org
- Listen and Repeat: 10 minutes a day of pause-and-repeat with short sentences
- Take an Interview: answer five brief factual questions about yourself, your studies and your routine, out loud, every day
- Record yourself once a day and listen for missing words and endings
Week 2: Stress, Rhythm and the First Frameworks (Days 8–14)
- Listen and Repeat: move to longer sentences; mark the chunks and stressed syllables before you repeat
- Shadow for 10 minutes a day, rotating North American, British, Australian and New Zealand speakers
- Take an Interview: practise experience, narration and opinion questions with the frameworks above, three a day
Week 3: Longer Sentences and Harder Questions (Days 15–21)
- Listen and Repeat: practise the longest sentences you can find, and repeat them after a single listen
- Take an Interview: add explanation and prediction questions, and answer four questions in a row, from factual to prediction
- Review your recordings for fillers, rushed speech and answers that stop after one sentence
Week 4: Full Run-Throughs (Days 22–30)
- Two or three times a week, do a full set: seven sentences to repeat, then four interview questions, without stopping in between
- Keep practising the rest of the test too: the adaptive Reading and Listening sections and the new Writing tasks
- On test day: listen to the whole prompt, start promptly, and speak clearly rather than quickly
Daily Practice Routine (25–30 minutes)
- 5 minutes: warm up by reading a short passage aloud, focusing on clear sounds and stress
- 10 minutes: Listen and Repeat drills or shadowing
- 10 minutes: three or four interview questions, recorded
- 5 minutes: listen back and note one thing to fix tomorrow
Practising the New Tasks with TalkDrill
TalkDrill has no TOEFL mode and does not give TOEFL scores. What it can give you is daily speaking practice for the skills the two tasks use:
- For Listen and Repeat: the free Echo Chamber game plays a model sentence, you say it back, and it scores how closely and how quickly you matched it. Read Aloud practice scores every word you read for pronunciation, and misread words come back highlighted.
- For Take an Interview: talk with an AI tutor by voice and ask it to interview you about your studies and campus life, or rehearse experience and opinion answers in practice interviews. The practice interview questions are written for job interviews, not for the TOEFL, but the skill of answering out loud in a structured way carries over, and each session ends with a scored report on fluency, grammar, vocabulary and pronunciation.
What Changed in 2026 (the Old Format)
Many books, videos and older guides, including earlier versions of this page, describe the TOEFL Speaking section as it was before January 2026. If you see any of the following, the material is about the old test.
Old Format vs. New Format
| Feature | Before January 2026 | Since January 2026 |
|---|---|---|
| Tasks | 4: Task 1 independent (opinion); Tasks 2–4 integrated (campus announcement, academic concept, lecture summary) | 11 items: 7 Listen and Repeat + 4 Take an Interview |
| Timing | 15 to 30 seconds of preparation, then 45 or 60 seconds to speak; about 16–17 minutes in total | Short, timed responses; about 8 minutes in total |
| Reading and listening inside Speaking | Yes: Tasks 2–4 asked you to summarise a passage and/or a lecture | No passages or lectures to summarise |
| Scoring | Each task rated 0–4, converted to a Speaking score of 0–30 | AI-scored, up to 5 points per item; Speaking score on the 1–6 scale |
| Total score | 0–120 | 1–6 overall, plus a comparable 0–120 score during a two-year transition |
What still carries over: the old Task 1 habit of stating a clear position with a reason and an example works well for the opinion questions in Take an Interview. Clear pronunciation, correct word stress, a calm pace and few fillers matter as much as ever.
What to drop: preparation-time timers, note-taking drills for reading-and-lecture summaries, and any template built around "the reading says… the professor says…". The integrated writing task has gone too; in Writing, only Write for an Academic Discussion survives from the old test.
Practise Listen and Repeat for Free
Echo Chamber plays a model sentence and scores how closely and how quickly you say it back. It is free, and it is the closest TalkDrill drill to the new Listen and Repeat task. TalkDrill has no TOEFL mode and gives no TOEFL scores.
Try Echo Chamber →